Identification of the Impact of GNSS Positioning on the Evaluation of

Informative Speed Adaptation

Jamal Raiyn

Computer Science Department, Al Qasemi Academic College, Israel

Keywords: GNSS Data, Positioning Performance, Autonomous Vehicle, ISA.

Abstract: Autonomous vehicles (AVs) are self-driving vehicles that operate and perform tasks under their own power.

They may possess features such as the capacity to sense environment, collect information, and manage

communications with other vehicles. Many autonomous vehicles in development use a combination of

cameras, various kinds of sensors, GPS, GNSS, radar, and LiDAR, with an on-board computer. These

technologies work together to map the vehicle’s position and its proximity to everything around it. To

estimate AV positioning, GNSS data are used. However, the quality of raw GNSS observables is affected by

a number of factors that originate from satellites, signal propagation, and receivers. The prevailing speed

limit is generally obtained by a real-time map matching process that requires positioning data based on a

GNSS and a digital map with up to date speed limit information. This paper focuses on the identification of

the impact of GNSS positioning error data on the evaluation of informative speed adaptation. It introduces a

new methodology for increasing the accuracy and reliability of positioning information, which is based on a

position error model. Applying the sensitivity analysis method to informative speed adaptation yields

interesting results which show that the performance of informative speed adaptation is positively affected by

minimizing positioning error.

1 INTRODUCTION

For most intelligent transport systems (ITSs), the

impact of the quality of the positioning information

on ITS user service-level performance cannot be

easily estimated. However, it can be of fundamental

importance for critical services, and therefore calls

for detailed analysis. Over the last few years, various

geo-positioning technologies have been used to

estimate the location of vehicles (Du et al. 2004;

Quddus et al., 2007), such as satellite-positioning

technologies (i.e. global navigation satellite systems

[GNSSs], and global positioning systems

[GPSs](Ramm and Schwieger, 2007), wi-fi

positioning systems, and cellular positioning

systems(Alger, 2014; Zandbergern, 2009). Some of

the methods for collecting data on road traffic flow

involve fixed-point modes, with high costs and

limited regional coverage, such as induction loops

and radar and video techniques (Fleming, 2001;

Groves, 2013). In contrast to these fixed-point

modes, we have introduced a system of floating data

management based on augmented GNSS-based

terminal positioning to improve the estimation of

vehicle location, for road ITSs (Raiyn, 2016). The

advantages of GNSS-based positioning are: accuracy

and low processing time complexity. The basic

operating principle of satellite navigation systems is

to calculate a user’s position from a GNSS signal.

However, the quality of raw GNSS measurements

(also called observables) is affected by several

factors that originate from satellites, signal

propagation, and receivers.

This paper is organized as follows: Section 1

gives an overview of the technology used to estimate

AV positioning; section 2 explains the collection of

raw GNSS data; section 3 describes the system

model; section 4 presents the method for identifying

GNSS positioning error; section 5 discusses the

results of the implemented informative speed

adaptation and section 6 concludes the paper.

2 GNSS POSITIONING DATA

In transportation, positioning can be monitored

everywhere, and the number of road transport

systems using positioning systems, for the most part

Raiyn, J.

Identiﬁcation of the Impact of GNSS Positioning on the Evaluation of Informative Speed Adaptation.

DOI: 10.5220/0007656903050311

In Proceedings of the 5th International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2019), pages 305-311

ISBN: 978-989-758-374-2

305

based upon GNSS, is almost infinite. The

requirements of these systems with respect to the

positioning information provided by the positioning

terminal can vary from decimetres to hundreds of

metres, depending on the application. Some of these

systems are critical in terms of the need for safety,

liability or security, and they depend on accurate and

reliable positioning information to function

effectively. For ITSs and LBSs, a standard in 2015

is a single-frequency GPS L1 receiver, with

increasing SBAS and dual-constellation (with

GLONASS) capabilities. LBS receivers

(smartphones) are also often assisted (A-GNSS) by

and even hybridized with positioning based on

communication networks. The ability to use multiple

GNSSs improves the accuracy, integrity and

availability of positioning, especially in urban areas.

There is a very large offer of different types of

GNSS receivers today, with highly variable

performances and costs. The performance of a

GNSS may be improved by data fusion, namely by

integrating sensor measurements, other positioning

means or a priori data such as digital data bases.

This data fusion brings enhancement at three levels:

(1) standard position, velocity and time (PVT) are

improved in terms of availability, accuracy, and

integrity; (2) additional information, such as attitude

angles, may be provided, and (3) the output rate is

increased by one or two orders of magnitude. One of

the most influential trends in GNSS is the use of

multiple systems to achieve better error mitigation

(e.g. multipath), resistance to interference and

positioning accuracy. Ground and satellite based

augmentation systems will also be used more in the

future to improved position accuracy and integrity.

Furthermore, in this project we will consider cyber

and information security in an augmented GNSS,

which may influence PVT.

The GNSS positioning principle relies on

trilateration by which an unknown point location

(receiver) is estimated using distance measurements

observed from known point locations

(satellites)(Groves, 2013). The basic observable of

the system is the travel time required for a signal to

propagate from the satellite to the receiver

multiplied by the speed of light to compute distance.

The receiver could then be located anywhere on the

surface of a sphere centered on the satellite with a

radius that equals this distance. The quality of raw

GNSS observables is affected by several factors

originating from satellites, signal propagation, and

receivers. The signal transmitted by a satellite

propagates through the atmosphere, where it is

subject to delays caused by the ionosphere and

troposphere. The effects of these delays are only

partially compensated for by global models in single

frequency receivers.

At the ground level, multipath, namely the

reception of signals reflected from objects like

buildings surrounding the receiver, can occur,

inducing one of the largest errors that is difficult to

model, as it is strongly depends on the receiver

environment. The worst situations are experienced

when only reflected signals are received (non-line-

of-sight signals, or (NLOS) signals, resulting in

pseudo-range errors of several tens of metres or

greater in extreme cases.

Finally, random errors are encountered at the

receiver level due to receiver thermal noise. The

receiver clock offset (much larger than that of the

satellite) does not create any error, since it is

considered as an unknown and is calculated together

with the position. The position error that results from

the measurement errors above, which is referred to

as dilution of precision (DOP) depends also on the

relative geometry between the receiver and the

satellites. Accuracy is maximized when the

directions to tracked satellites are more uniformly

spread around the receiver.

The main task of a GNSS is to provide

localization and time synchronization services.

There are multiple GNSS systems available. The

most well-known one is the global positioning

system (GPS). GPS data is transmitted via

coarse/acquisition (C/A) code, which consists of

unencrypted navigation data. The encrypted

(military) signal is called the precision-code, which

is also broadcast by every satellite. It has it is own

PRN codes, but it is in the order of 1012 bits long.

When locked onto the signal, the receiver receives

the Y code, which is an encrypted signal with an

unspecified W code (Loukas et al., 2013;

Humphreys, 2013). Only authorized users can

decipher this. In later GPS satellites, extra features

were added (Radoslav et al., 2014; Uma and

Padmavathi, 2013). There are several methods of

augmenting GNSS data to get better estimates of

location. Three of these are satellite-based

augmentation systems (SBASs), assisted-GPS, and

differential-GPS. SBASs were the first type to be

developed; these systems are commonly used in

airplanes, for critical phases such as the landing

phase. They consist of a few satellites and many

ground stations. A SBAS covers a certain GNSS for

a specific area, and for every GNSS, accuracy

depends heavily on, and is influenced by external

factors (Grove, 2013). These factors affect not only

GNSS applications, but also every other wireless

VEHITS 2019 - 5th International Conference on Vehicle Technology and Intelligent Transport Systems

306

transmission application. Furthermore, the satellites

orbit the earth at a height of approximate 20.000 km,

and at this height, signals can be affected in many

ways. According to (Grove, 2017), “space weather”,

and solar conditions affect signals too. A GNSS

requires exact timing on the order of nanoseconds to

determine position, but when, a satellite signal

reaches earth, it can reflect from buildings and other

objects, causing an increase in travel time and

influencing measurements.

3 SYSTEM MODEL

In this model, the urban road is divided into sections,

and each road section is assigned an intelligent

vehicle agent (IVA). The goal of the IVA is to

manage the adaptive traffic and to serve the

autonomous vehicle.

The functionality of the system model works on

three levels: the performance of the GNSS- based

positioning terminal, and the performance of the

system at the service error level. Autonomous

vehicles (AVs) are sometimes referred to as

driverless vehicles or self-driving vehicles. They

offer many advantages and are expected to appear

on the commercial market by 2030. Comfort is one

obvious advantage, but in the current society, the

practical advantages of AVs become clearer every

day. Due to increased congestion on roads,

productivity decreases and money is wasted on fuel

and time. Cooperative AVs (Jawhar et al., 2013)

enhance traffic flow, and in regard to road safety,

smart vehicles are likely to decrease the number of

injuries and fatalities. AVs collect information from

the environment during their activity. However,

these autonomous technologies are subject to

malicious input, and are lacking in security. From a

security-by-design perspective this is wrong,

because a decision made by an AV is only as good

as the sensors can perceive. A faulty observation can

lead to dangerous situations.

4 METHODOLOGY

The performance of any system can be characterized

with respect to different features, or characteristics.

For ITS and mobility applications, the most relevant

performance features are availability, accuracy and

integrity. These features can be relevant for different

outputs of the terminal, such as horizontal position,

altitude, longitudinal speed, etc. As a result, a large

Figure 1: System model.

number of combinations of features and outputs can

be assessed during the performance characterization

of a positioning terminal. Each characterization

process calls for a performance metric (or indicator).

A performance metric involves a precise definition

of the means of measuring a given performance

feature of a given system output. A performance

metric is necessarily associated with a standardized

test protocol defining the test procedures and means

(test vehicle, simulator, record and replay

equipment, etc.), the test scenario and the way the

test results will be transformed into indicators. The

test scenario should precisely define the set-up

conditions of the terminal (in particular of the GNSS

receiver antenna), the trajectory, the sample size, the

environmental conditions, the interference

conditions, so forth. It should reflect as faithfully as

possible the real operational conditions under which

the ITS system using the positioning terminal will

operate.

 Availability in ITS

Availability refers to the percentage of time

during which the output of the positioning terminal

is available. This feature can be defined in many

different ways according to the needs of the

application. In general, availability refers to the

percentage of the measurement epochs (time

periods) when the considered output is delivered

with the required performance. A more

straightforward metric would simply be the

percentage of the measurement epochs when the

considered output is delivered by the terminal,

regardless of its quality.

 Accuracy in ITS

This feature refers to the statistical

representation of the merit of position error, velocity

error or speed error. Accuracy metrics have to be

constructed based on the statistical distribution of

errors.

intelligent cooperative traffic control algorithm

V2V

V2I

section i section j

critical zone

IVAi IVAj IVAn

section n

V2IoT

Identiﬁcation of the Impact of GNSS Positioning on the Evaluation of Informative Speed Adaptation

307

 Integrity in ITS

The concept of integrity was introduced in civil

aviation to measure performance affecting safety, for

example, executing a safe landing. It is not enough

that errors be small on average (accuracy); they must

remain small for every landing (integrity). Given the

focus of this document, dedicated to road transport

and mobility, the definitions for integrity are

inspired by, but significantly simpler than, the

definitions used by the civil aviation community.

High accuracy does not mean high integrity.

4.1 Positioning Error Modelling (PEM)

Position error modeling is mandatory for a

sensitivity analysis that checks for the compatibility

of a given position terminal with a given application

algorithm in a given environment. In view of what

we have mentioned in previous sections, it is

challenging to model errors at the level of raw

measurements (or observables) and to propagate this

model through the navigation algorithm, which most

of the time is unknown and is always non-linear.

The most efficient method is to identify a model that

captures real errors observed at the position error

modeling level as closely as possible. The proposed

methodology applies to various environments, which

calls for a variety of models.

Figure 2 illustrates the sensitivity analysis, which is

based on field tests of a GNSS-based positioning

terminal (GBPT) carried out under real conditions to

identify a positioning error model.

Figure 2: Operational scenario.

4.2 Quality of GNSS Data

AVs may receive data from various sources. The

central monitor has the task of measuring the quality

of received data in the AV network. There are two

kinds of data: data for travel forecasting (Broggi et

al., 2013; Ramm and Schwieger, 2007) and data for

vehicle positioning estimation (Alger, 2014). In

general, data quality is defined as data that is

suitable for use (Raiyn, 2017). In AVs data quality

involves delivering the right data to the right user at

the right time and correct data enables the making of

reliable, and accurate decisions. Low quality data

can lead to traffic congestion and collision. In

general, there are three main approaches that use

performance metrics to test the global performance

of a GNSS-based positioning terminal: filed tests,

lab tests, and record and replay tests. This section

discusses the impact of these approaches.

 Field Tests

Field tests were performed by SaPPART to

collect data. The car used was equipped with uBlox

and smartphones. Figure 3 illustrates the data

collection in North Paris. The data are incomplete

due to urban noise produced by network tunnels.

Figure 3: Field tests.

 Lab Tests

Simulations were carried out using MATLAB,

due to the fact that the observations of the traffic

were incomplete and were a noisy function of the

unobservable state process which can be observed

only through noise measurements. In cases where

the system received incomplete GNSS data as

illustrated Figure 4, the system combined historical

GNSS data with map matching to supply the missing

data.

Figure 4: Lab tests.

VEHITS 2019 - 5th International Conference on Vehicle Technology and Intelligent Transport Systems

308

The newly received data set was compared to the

available data set. Innovative data were

characterized by statistical measurement as

illustrated in Figure 5.

Figure 5: Data analysis based on SEM.

4.3 Algorithm Description

The algorithm describes the message exchanges

between an intelligent vehicle agent (IVA) and an

AV. The IVA updates the AV about the urban road

traffic through a communications protocol based on

message exchanges.

Algorithm I: send message

for all AV

At ts

leader sends its current acceleration and speed

At ts

j-1,

receives its current speed v

and acceleration

At ts

, leader gets from IAV

End for

Algorithm II: received message

for all AV

At ts

autonomous vehicle AV

receives the acceleration and

speed v

from the leader

At ts

j-1,

receives the acceleration from its precedent vehicle

At ts

transmits its current acceleration to its follower

End for

Algorithm III: compute forecasting speed

for all AV

At ts

leader computes its forecasting speed

At ts

leader sends its next acceleration and speed

forecasting

At ts

leader gets from IVA

computed next acceleration

to use at the next updating cycle

End for

Algorithm IV: follower strategy

for all AV

At ts

, AV

receives the acceleration and speed

from the leader

At tsj-1, AV

computes its predicted speed based on its

current acceleration

At tsj-1, AV

compute its next acceleration

At tsj, AV

transmits its next acceleration to its follower

End for

4.4 Informative Speed Adaption

Informative speed adaptation (ISA) uses the

exponential moving average (EMA) scheme for

travel speed forecasting. The algorithm forecasts

travel observations based on the EMA for the

designated urban road. Its procedure is as follows.

4.4.1 Short-Term Forecasting based on

Historical Information

The historical database is a collection of past travel

observations of the system. The exponential

smoothing forecasting method gives unequal weight

to the observed time series. This is accomplished by

using one or more smoothing parameters, which

determine how much weight is given to each

observation. The major advantage of the exponential

smoothing method is that it provides good forecasts

for a wide variety of applications. In addition, the

data storage and computing requirements are

minimal, which makes exponential smoothing

suitable for real-time forecasting.

),()1(),(),1( ktttktttkttt





(1)

where

10 



, tt

(t, k) is the actual travel time

in section k at time t, and tt

(t, k) is the historical

travel time in section k at time t.

4.4.2 Short- Term Forecasting based on RT

Information

The occurrence of abnormal conditions in traffic

flow travel information decreases of forecasts based

on historical information and may increase the

complexity of the forecasting of unusual incidents.

The forecasting model based on real-time

information gives a little weight to historical

information and great weight to real-time

observations.

)),(),((),1(),1( ktttktttktttkttt

HMH





(2)

where

10 



4.5 Description of the Positioning Error

The ISA application is based on geo-objects

specified by latitude and longitude. Positioning

errors and angle errors were cloned and these clones

were combined to create “clone trajectories”. The

reference trajectory was used, and the clone points

were computed by adding cloned radius and angles

into reference points as illustrated in figure 6.

Identiﬁcation of the Impact of GNSS Positioning on the Evaluation of Informative Speed Adaptation

309

cloned

= x

ref

+sin(yaw

ref

–Angle)*Radius (3)

cloned

ref

+cos(yaw

ref

–Angle)*Radius (4)

Figure 6: Positioning error model.

5 IMPLEMENTATION

This section describes informative speed adaptation

(ISA). The algorithm introduced here and used for

informative speed adaptation (ISA) is also known as

an intelligent speed limiter, or intelligent speed

assistance, or an intelligent speed alerting or

intelligent speed authority. The purpose of ISA is to

mitigate speeding, (i.e., drivers’ travelling at speeds

above the legal speed limit). This is accomplished by

informing or alerting the driver, or even by taking

control of the vehicle, depending on the system

design. The prevailing speed limit is generally

obtained through a real-time map-matching process

that requires localization via GNSS and a digital

map with up-to-date speed limit information.

Applying a GNSS to calculate travel time has

proved to be effective in terms of accuracy. In this

case, GNSS data are managed to reduce traffic

congestion and road accidents.

The road is divided into sections as illustrated in

Figure 7. For each section the speed has been

forecast based on real-time travel observations. The

IVA system updates each AV on the speed limit in

each section. An intelligent vehicle agent (IVA)

updates each AV on the travel speed in its road

section. Figure 8 illustrates the impact of GNSS

positioning error on the evaluation of ISA. In

sections where the GNSS positioning error is high,

the traffic congestion is greater (red color). Figure 9

shows that the performance of ISA is improved by

positioning error correction. Figure 10 illustrates the

short term travel forecasts for road sections. The

travel forecasting is based on exponential moving

average with the optimization of alpha and gamma.

Figure 11 illustrates the evaluation of the

forecast scheme EMA. The root mean square error

increases in sections where GNSS positioning error

is high.

Figure 7: Road partitioning.

Figure 8: Travel speed.

Figure 9: Positioning correction.

Figure 10: Mean travel speed prediction.

VEHITS 2019 - 5th International Conference on Vehicle Technology and Intelligent Transport Systems

310

Figure 11: Root mean square error.

6 CONCLUSION AND FUTURE

WORK

In general, schemes that have been proposed involve

forms of centralized management, which may

achieve near- optimum performance in whole

systems in regard to maximum capacity. However,

as the number of vehicles increases, centralized

computation may become mathematically

intractable. Such schemes are also impractical when

traffic varies significantly and creates difficulties for

measuring actual conditions. The major

disadvantage of centralized management is the

occurrence of deadlock, which causes the whole

system to collapse. To reduce the overload in

computational management time, we consider the

management of urban road traffic in distributed

form, and for this, we propose the use of

decentralized management. The proposed approach

is based on intelligent vehicle agent techniques; it

aims to reduce several types of traffic congestion

and their effects, such as delays, waiting time, driver

stress, air and noise pollution, and economic costs.

Informative speed adaptation is used to update AVs

in order to determine the shortest route from the

source to the destination node based on short- term

forecasting. The update phase is used to inform the

AVs about new events during the trip, and the

updated information is used to reduce road traffic

congestion. In keeping with the updated information,

vehicles are allocated the appropriate road sections,

and drivers can select new sections with low traffic

congestion.

REFERENCES

Alger, M., 2014. Real-Time Traffic Monitoring Using

Mobile Phone Data, Proceedings of 49th European

Study European Study Group with Industry, Oxford,

United Kingdom.

Broggi, A. et al., 2013. Extensive Tests of Autonomous

Driving Technologies, IEEE Transactions on

Intelligent Transportation Systems 14.3. pp. 1403–

1415.

Du, J., Masters, J. , Barth, M., 2004. Lane-Level

Positioning for In-Vehicle Navigation and Automated

Vehicle Location (AVL) Systems. 2004 IEEE

Intelligent Transportation Systems Conference, pp. 1–

Fleming, W. J. J. 2001. Overview of automotive sensors.

IEEE Sensors Journal 1.4, pp.

296–308.

Groves, P.D. 2011. Shadow Matching: A New GNSS

Positioning Technique for Urban Canyons. The

Journal of Navigation, 64, pp.417–430.

Groves. D.P., 2013. Principles of GNSS, inertial, and

multisensor integrated navigation systems. Artech

house.

Loukas, G., Gan, D., Vuong, T. 2013. A Review of Cyber

Threats and Defence Approaches in Emergency

Management. In: Future Internet 5.2, pp. 205–236.

Humphreys,T.E. 2013. Detection Strategy for

Cryptographic GNSS Anti-Spoofing. In: IEEE

Transactions on Aerospace and Electronic Systems

49.2 . pp. 1073–1090.

Jawhar, I., Mohamed,N., Usmani, H. 2013. An Overview

of Inter-Vehicular Communication Systems, Protocols

and Middleware. Journal of Networks 8.12 (Dec.

2013), pp. 2749–2761.

Quddus, M.A. Ochieng, W.Y,. Noland, R.B.. 2007.

Current map-matching algorithms for transport

applications: State-of-the art and future research

directions, Transportation Research Part C 15. pp.

312–328.

Ramm, K. and Schwieger, V., 2007. Mobile positioning

for traffic state acquisition, J. Location Serv.; 1(2).

pp. 133–144.

Radoslav , I., Pajic, M. Lee., I. 2014. Attack-resilient

sensor fusion. Proceedings of the conference on

Design, Automation & Test in Europe. European

Design and Automation Association., p. 54.

Raiyn,J., 2016. Speed Adaptation in Urban Road Network

Management, Transport and Telecommunication, 17,

2. pp. 11-121.

Raiyn, J. 2017. Road traffic congestion management based

on search allocation approach, Transport and

Telecommunication. 18. (1). pp. 25-33.

Uma, M. Padmavathi,G. 2013. A Survey on Various

Cyber Attacks and their Classification, International

Journal of Network Security, vol. 15, no. 6, , pp. 391-

397.

Zandbergen, A.P., 2009. Accuracy of iPhone Locations: A

Comparison of Assisted GPS, WiFi and Cellular

Positioning, Transactions in GIS, 13(1), pp. 5-16.

Identiﬁcation of the Impact of GNSS Positioning on the Evaluation of Informative Speed Adaptation

311