satellite navigation & sensing: a match made in...

University of ColoradoBoulder

Satellite Navigation & Sensing: A Match Made in Heaven

Jade MortonUniversity of Colorado Boulder

Satellite Navigation and Sensing Lab

Stanford PNT Symposium 10/30/2019 1


Society’s Increasing Dependence on GNSS

2


Satellite Navigation ModernizationGPS GLONASS BeidouGalileo QZSSIRNSS

SBAS

1 Civil Signal 6 Civil Signals, 3 frequencies

2023: >160 navigation satellites> 400 signals

Now: 25~35 satellites in view

PNT: Need 4 SVs in view

Position, Navigation, Timing3

, & Sensing (PNTS)


GNSS and Propagation Environment

Multipath reflection

Ionosphere

TroposphereGNSSRadio Occultation GNSS

Reflectometry

Interference

Propagation effects: • Absorption• Bending• Delay • Reflection• Scattering

What GNSS has to offer: • Free• Large number• Distributed • Well-defined • Diversity

The Last 5% of the Trip


GNSS Receiver

5

• Foliage• Indoor/urban• Interference• Ionosphere• Multipath• TropospherePlatform

• Aircraft• Spacecraft• UAV• Ground

Sensingand

SituationAwareness

Characterization

InformationRetrieval

PNTunder

Challenging Conditions

Impac

t

Mitigati

on

Techniques


Classification of GNSS Receiver Challenges

6

• Indoor/Urban• Interference• Foliage• Multipath• Ionosphere• Troposphere

Weak Signal

• Aircraft• Spacecraft• UAV

Dynamic Platform

ScintillationAmplitude Fluctuation

Phase Fluctuation


Ionosphere

Water vapor

Volcano plumes

Scintillation: amplitude, frequency, phase

fluctuation

What Do Ionosphere, Troposphere, Foliage, and Multipath Have In Common?

7

Low-elevation satellites suffer more but contain more information


A Simple Simulation: Sum of 2 Rays

Singapore Multi-GNSS Update 8

!"!#

= 0.95)" − )#= -1Hz

!"!#

= 1.1)" − )#= 1Hz


Example: Ionosphere Scintillation

Singapore Multi-GNSS Update 9

L1L2

L5


10

Mountain Top Data Collection ExperimentHaleakala Summit: 10,023 ft


11

Elevation

Troposphere Scintillation & OceanMultipath Reflections

5/6/2017HaleakalaMaui

L1L2

L5


12

A Closer View at 21-22 min

L1

L2

L5


13

L1

L2L5

Multipath On L5 Disappears As SV Elevation Increases


Application Objectives

14

PNTunder

Challenging Conditions

Sensingand

SituationAwareness

Disturbance signaturesRange accuracy

RobustnessMaximize MaximizeMaximize Don’t CareMinimize Maximize


Two Challenges in GNSS Sensing Applications

15

1. Data availability

2. Data quality

• RX lose lock in challenging propagation environments

• Discontinuity in measurements• Noise

• Need more robust RX architecture

• Need better filter design and better aiding mechanisms


Multi-Domain GNSS Receiver Processing

• Adaptive tracking• Inter-frequency aiding• Vector processing• Semi-open loop• Open loop• Array processing

Parameter optimizationFrequency diversitySignal spatial diversity

Receiver diversityEnvironment informationTemporal diversity

16


Commercial ISM Receiver

RF Front End 1

RF Front End 2

RF Front End N

Space Weather Events

Data Collection and Control Server

Space Weather Event Monitoring & Trigger

Software

Circular Buffer

Circular Buffer

Circular Buffer

Data

Storage

VPN

Data Center at Home Institution

Inte

rnet

Receiver Signal Processing Algorithm Library

Ionosphere Studies

ImproveGNSSReceiver

Event-drivenDataAcquisitionSystem

18


Software-Defined GNSS NetworkToolik Lake

Poker FlatGakona Greenland 1

Andoya, Norway

Hong Kong

Singapore

McMurdo

Chile

Peru

Puerto Rico

Svalbard, Norway

EthiopiaAscension Island

Hawaii

Boulder

Ecuador

India

South Korea

Resolute Bay

Greenland 2

Funded, to be deployed by 2020

Existing


GNSS Event-Driven Data Acquisition System (EDAS)


Three Application Examples

• Ionosphere: – Ground-based GNSS Network for Space Weather Monitoring

• Troposphere: – Radio Occultation Detection of Planetary Boundary Layer

• Surface: – Coherent Reflection for Precise Ocean Altimetry

20


September 7-8, 2017 Storm Impact on PPP

Storm Commencement


Spatial-Temporal Machine Learning Architecture

⋯

Input"#$% "#$%&' "#Forecast

"#&(

Spatial Fusion

Fused Featuresat "#$%

Spatial Fusion Spatial Fusion

Fused Featuresat "#$%&'

Fused Featuresat "#

Temporal Fusion

PNT Impact

Forecast

Machine Learning Framework

Forecast Model

⋯

⋯ External Features (e.g. Sunspot #, F10.7, etc.)


GNSS Radio Occultation

24

cosmic.ucar.edu


Large Errors in Lower Troposphere

25

Altit

ude

(km

)

Refractivity Error


Mountaintop Radio Occultation Experiment Concept

GNSS LEO

High gain antenna

Multipath reflection/scattering from ocean surface

26


RO Application Example: Planetary Boundary Layer Height

27

Carrier phase à bending angle à refractive index à refractivity gradient à minimum gradient à PBLH

19 20 21 22 23 24 25 26 27 28 29 30 31 32Time (min)

Sig

nal I

nten

sity

(dB

)

40

20

PBL


28

Dense Data Samples


CYGNSS: A NASA GNSS Reflectometry Mission

29

Zenith Antenna GPS L1 (RHCP)

2 Nadir Antenna (LHCP, up to 14dB Gain)


Machine Learning for Wind Speed Retrieval

University of Michigan 3/28/2019 30

2 Challenges:• High wind speed• Wind direction


Coherent vs. Non-coherent Reflections

31


CYGNSS Raw IF Data Coverage

136 DatasetsAug 2017 – Apr 2019

32

Processed by CYGNSSRaw IF (Gain<0dB)Raw IF (Gain>0dB)


Rain Forest

Bolivia

1/28/2018




Open Ocean

Java Sea

2/7/2018


Java Sea 2019-01-10 01:52:44UT SPIRE SVN 86 & GPS 15


Conclusions• Satellite navigation applications go far beyond PNT.• Navigation satellites offer signals-of-opportunity for a wide range

of high impact sensing applications.– Low cost, low power, small volume, distributed– Ionosphere, troposphere, land cover, ocean surface, urban area, EM

• Synergy between satellite navigation and sensing:– Sensing capability improves navigation solutions– Navigation signals enable sensing

• Challenges and opportunities– Sensing on dynamic platform, computation resources, data assimilation

satellite navigation & sensing: a match made in...

Documents