Impact of GLONASS on receiver design FDMA versus CDMA Frank Boon NiN Workshop 27/03/09
Dec 20, 2014
Impact of GLONASSon receiver design
FDMA versus CDMA
Frank BoonNiN Workshop 27/03/09
March 27, 2009 © Septentrio 2
Background
Prototyped GPS+GLONASS PolaRx2 based on 4 Proprietary RF Chips
not commercialized : poor GLONASS constellation
served as engineering lessons platform
GLONASS ‘resurrection’ increased demand
AsteRx2 designed as GPS+GLO+GAL from start
AsteRx2 serves as geodetic OEM GPS+GLO receiver in many products
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 3
Contents
Background
FDMA in detail
GLONASS Impact on: RF Design
Satellite Acquisition
Tracking
Positioning Algorithms
Receiver Compatibility
Summary
Outlook
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 4
FDMA in detail
Satellite identification division multiple access
GPS uses digital identification : one central carrier
GLONASS uses frequency identification
one PRN-code for all GLONASS satellites
12 individual frequency carriers in-view k : Frequency Number, Fn
more k-values defined
k=0 k=+1 k=+2k=-2 k=-1
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 5
FDMA in detail
Consequences of FDMA Requires wider band-pass filter
Individual filter response per carrier
Fn-SlotNR relation not defined / guaranteed SlotNR important : antipodal system, Fn reused
users accustomed to PRN, SlotNR feels like PRN
recently 2 SV-combinations changed Fn without announcement.
Individual wavelength for each satellite influence on RTK,PPP algorithms
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 6
Impact on RF design
Current GNSS spectrum
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 7
Impact on RF design
band-pass filter choices : individual filters
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 8
Impact on RF design
band-pass filter choices : combined GPS+GLO filters
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 9
Impact on RF design
Individual carriers per SV each GLONASS satellite excites different part of RF
Individual group delay and phase shift per SV pseudoranges and carrier phases will be biased
frequency [Hz]
ph
ase
dela
y [
rad
]
Filter Frequency Response
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 10
Impact on RF design
RF quality determines measurement quality linearize group delay : simplifies calibration
reduce slope of filter response
ensure temperature stability
bias around cm-level
Combined GPS+GLO RF equal mean RF group delay for GPS+GLONASS
GPS enables calibrate GLONASS hardware delays
improved temperature stability & coherency
frequency [Hz]
ph
ase d
ela
y [
rad
]
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 11
Impact on Satellite Acquisition
Previously cumbersome... GLONASS-M significant improvement
Fn and SlotNR relation not defined upfront
GLONASS : relation only defined in almanac
GLONASS-M : both in almanac & ephemeris
Cold Boot : acquire on Fn first GLONASS : wait for almanac before reporting SLot
GLONASS-M : directly available
RINEX and RTCM data SlotNR based
Still one aging non-M GLONASS satellite...
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 12
Impact on Tracking
implementation of receiver Clock Jumps odd Fn can violate integer number wavelengths
1 msec jump shifts c*10-3 / l = f*10-3 wavelenghts
spacing of 0.5625 MHz leads to 562.5 wavelengths
2 msec jumps only, or adjust phase measurements
non FDMA related P-code not encrypted but content unknown
tracking is based on assumptions of content
during past year three changes have happened requires firmware updates (brand indepedent)
use L2CA instead (GLONASS-M only)
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 13
Impact on Positioning Algorithms
PZ90-2 solved difference with WGS-84
Cumbersome Ephemeris format
Orbit prediction Quality still lacking
GPS-GLONASS time difference (MOSCOW UTC) varies over time, observed 100 nanosec / month
individual wavelengths adapt RTK and PPP algorithms
use abstract carrier phases : non-integer
pair closely space Fn : large phase residuals
RTCM/CMR generation requires receiver clock bias wrt GLONASS
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 14
Inter Receiver Compatibility
RTK based on double differences affected by mixing receiver brands or generations
Different HW design, different group delays RF implementations vary per receiver brand
DD-biases depend on included Fn combinations
Calibrate with Initial GPS only RTK solution determine phase shift and group delay
determination for each Fn single difference
monitor time behavior to compensate T drift
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 15
Summary
FDMA has a significant impact on design hardware : RF quality required
firmware : common firmware blocks not possible
Not relevant for end-user if properly designed
GLONASS-M is a step forward direct link between SlotNR and Fn
L2CA is future proof, avoid P accuracy degradation acceptable
GLONASS-K significant leap looking forward to CDMA
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 16
Outlook
Not meant as negative towards GLONASS great constellation to support GPS until Galileo
arrives
GPS+GLO improves DOP, height component
GPS+GLO improves availability
GPS+GLO improves integrity (redundancy)
When Galileo arrives GLONASS will modernize to CDMA
L3 overlaps with E5 : great DualFreq GPS+GLO+GAL
fantastic quadruple constellation opportunities next workshop on what will Compass (Beidou-2) bring...
GLONASS workshop Amsterdam
March 27, 2009 © Septentrio 17
Outlook
Thanks for listening.
GLONASS workshop Amsterdam