International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 1

Number1

March 2007

DGNSS Re-Capitalization

M. Pattinson & M. Dumville

Nottingham Scientific Ltd, Nottingham, UK

N. Ward

General Lighthouse Authorities, UK & Ireland

ABSTRACT: The General Lighthouse Authorities (UK & Ireland) DGPS service came into operation in 1998.

In common with other maritime DGPS services the equipment will need replacement over the next few years,

in fact computing and communications equipment have already been replaced. Replacement of existing

hardware with similar, dedicated Reference Stations and Integrity Monitors (RSIM) is the baseline option and

will form a fallback plan if other options prove not to be feasible. However, the choice of suppliers is limited

and once chosen, it would be difficult to diversify. Three other options can be identified: software RSIM,

Virtual Reference Station (VRS) and integration with Satellite Based Augmentation Systems (SBAS). The

software RSIM option draws on the experience of the United States Coast Guard. The VRS and SBAS

integration options are treated as potential alternatives to an onsite hardware or software RSIM, but the

possibilities of combining either or both with the software RSIM are also considered. All options took into

account the need for validation of system performance. This paper draws conclusions about feasibility,

performance, risks and costs of the different options and makes recommendations on the course to adopt.

1 INTRODUCTION

1.1 Purpose and Scope

The General Lighthouse Authorities (GLA) DGPS

service came into operation in 1998. The DGPS

equipment has a typical operational life of 5-10 years

and the computers and communications equipment

have already been replaced. The Reference Stations

and Integrity Monitors (RSIM) currently in use are

no longer in production and may not be supported

beyond 2008. Also, the communications between

RSIM and control station need to be upgraded as

they are currently based on proprietary protocols of

the RSIM supplier. Furthermore, the present GLA

DGPS system will not meet all the requirements set

out in IMO Resolution A.915(22) for Future GNSS

(IMO 2001). Four routes have been considered:

− Replacement of the existing hardware RSIM;

− Replacement with RSIM implemented in soft-

ware;

− Virtual Reference Station (VRS) approach, using

an existing network;

− Integration with Satellite Based Augmentation

Systems (EGNOS).

In order to investigate these options and provide

recommendations, NSL has carried out a study on

behalf of the GLA, the results of which are presented

in this paper.

For each route, several options for integration will

be analysed. All options will take into account the

need for validation of system performance. The pos-

sibility of building in the necessary infrastructure,

data processing, analysis and reporting will be

considered.

1.2 Study Logic

The study itself was divided into the following tasks:

− Concept Evaluation: Activities in this task will

propose and evaluate a concept for the incorpora-

tion of the “selected option” into the DGNSS ser-

vice in order to meet the requirements set out in

IMO Resolution A.915(22) (IMO 2001). The task

will define the physical components and op-

erating processes as well as the monitoring,

control and deployment concepts.

− Technical Analysis: This task will provide a base-

line definition of the overall functionality and

architecture of the proposed DGNSS service

options, as well as identifying the validation

infrastructure and tools will be required to

monitor the performance of the service.

− Performance Assessment: The performance of the

DGNSS service architecture will be assessed and

analysed by simulations using NSL’s NEMO tool.

The main performance criteria to be analysed will

be the ability to deliver the necessary levels of

accuracy, integrity, availability and continuity

performance across the GLA coverage zone.

− Critical Items and Risks: This task will identify

critical functions, interfaces and core elements

that are essential to the DGNSS service in order

to meet the requirements. The task will identify

associated technical and operational risks that

may impact the delivery or performance of the

DGNSS service.

− Economic Analysis: The activity will provide

information on the incremental costs for

development, deployment and operation of the

DGNSS service that can be attributed to the

inclusion of the option. Costs will be sourced

from suppliers and through discussions with

European bodies.

− Recommendations: The final activity will be to

present a summary and conclusions of the tasks

and to provide a series of recommendations for

implementation.

2 CONCEPT EVALUATION

2.1 Introduction

As the developments in GNSS will be incremental, it

is unlikely that a single option can be identified that

can be implemented now and will cover all future

requirements set out in IMO Resolution A.915(22)

(IMO 2001). In fact, the timeline of DGPS re-

capitalization makes it clear that it will be a two

stage process. Therefore each considered option will

be assessed in terms of both backward compatibility

with legacy users and flexibility for modification

/upgrade as new signals/systems come on line.

In addition, it may not be most efficient to have a

single solution to cater for both global and regional

solutions and local high accuracy services. Therefore

the assessment of options for local high accuracy

operations (possibility of 0.1 m accuracy) is separated

from the section detailing possible options for

providing global and regional services (10 m and 1

m accuracy).

2.2 Global/Regional Solutions

To provide global/regional solutions, a number of

options were identified. After initial analysis, the

following options were chosen as the best to take

forward for further study:

− Procurement of upgraded HW RSIM: This

option involves a direct replacement of the

existing HW with upgraded COTS HW RSIM,

along with upgraded transmitters;

− Procurement of upgraded HW + SW RSIM:

This option involves implementing the main

RSIM functionality in SW that is external to the

GNSS receivers. This allows for greater

flexibility than the baseline option as it should be

easier to upgrade SW to for new message types

and signals rather than having to procure new

HW;

− OSNET VRS (DGPS): Instead of operating

separate RSIM sites, in this option a central

processing facility uses data from 3rd party

network to compute corrections for all virtual

reference stations, that are then sent to the

transmitters;

Table 1. Summary of global/regional re-capitalization options

DGNSS Option

Backward

compatibility

Maturity of

Technology

Flexibility for

message upgrade

Flexibility for

future GNSS

signals

Possible

Reduction in

Infrastructure

Procurement of upgraded

HW SIM

Yes High Medium Medium No

Procurement of upgraded

HW + SW RSIM

Yes Medium High Medium No

OSNET VRS (DGPS)

Yes

Medium

High

Low (*)

Yes

SW Radio receiver + SW

RSIM

Yes Low High High No

EGNOS RTCM

Yes

Low

High

Low (*)

Some

EGNOS RTCA

Low

High

Low (*)

Some

(*) – Any upgrades are dependent on 3

party evolution plan and are therefore not controllable by the DGNSS service provider

Table 2. Summary of local high accuracy options

DGNSS

Option

Maturity of

Technology

Infrastruct

ure costs

Coverage

Existing

standards

Reliance on

party

OSNET

VRS (RTK)

Medium Medium

Good within

network – up to

20 km outside

Yes Part

EGNOS

WARTK

Low Low Good No Yes

− SW Radio + SW RSIM: This option is similar in

architecture to the Procurement of upgraded HW

with SW RSIM option. However, instead of a

HW GNSS receiver and beacon receiver at the

RSIM sites, SW Radio technology is used;

− EGNOS RTCM: In this option the RTCM

correction messages are computed using the

EGNOS correction messages, obtained via

EDAS. Therefore the reference stations simply

retrieve the EGNOS messages and morph them

into RTCM format for broadcast. The Integrity

Monitor architecture and functionality is

unchanged;

− EGNOS RTCA: In this option the EGNOS

correction messages are obtained via EDAS and

simply re-broadcast via the DGPS transmitter

network so that users in high latitudes, or in areas

where the view to the EGNOS GEO is restricted,

can enhance their solution using the EGNOS

RTCA corrections.

Table 1 summarises the main advantages and

disadvantages of the identified re-capitalization

options.

2.3 Local High Accuracy Solutions

To provide local high accuracy solutions, a number

of options were identified. After initial analysis, the

following options were chosen as the best to take

forward for further study:

− OSNET VRS (RTK): The architecture of this

option is very similar to the OSNET VRS code

option, except that RTK messages rather DGPS

messages are produced for virtual reference

stations;

− EGNOS WARTK: In this option, instead of

operating separate RSIM sites, a central

processing facility computes corrections for the

whole area that are then communicated to the

transmitters. The input data used to compute the

corrections is obtained from the EGNOS RIMS

via EDAS.

The following table summarises the main

advantages and disadvantages of the identified re-

capitalization options.

2.4 Conclusions of Concept Analysis

One of the main drivers for re-capitalization will be

the maturity of the technology and the possible dates

at which such a service could become operational.

Based on best estimates for operational dates, a

timeline illustrating DGPS re-capitalization schedule

and possible dates when the different options may be

available is shown in the Figure 1.

Procurement of upgraded HW EGNOS RTCM

Procurement of upgraded HW + SW RSIM EGNOS RTCA

RTK EGNOS WARTK

OSNET VRS (code and carrier)

PPP Galileo CS

EGNOS GEO

EGNOS DGNSSSW Receiver + SW RSIM

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Specificatio

Tendering

Installation

operations

1st generation recapitalisation program

Specificatio

Tendering

Installation

operations

L1 DGPS

2nd generation recapitalisation program

L1/L5 DGNSS

Fig. 1. Timeline of DGPS re-capitalization an availability of

options

Therefore there are 4 options that are considered

for 1

generation re-capitalization and a further 4

that are considered for 2

generation:

− 1st Generation Options

− Procurement of upgraded HW RSIM

− Procurement of upgraded HW + SW RSIM

− OSNET VRS (DGPS)

− OSNET VRS (RTK)

− New 2nd Generation Options

− SW Radio + SW RSIM

− EGNOS RTCM

− EGNOS RTCA

− EGNOS WARTK

NB It should be noted that those options

considered as potential 1

generation options can

also be considered for upgrade to 2

generation.

3 PERFORMANCE ASSESSMENT

3.1 Methodology

The performance assessment has been carried out

using NSL’s NEMO SW tool. NEMO is a Service

Volume Simulator (SVS), and is a fully interactive

PC-based Windows application. It provides a

flexible and extendible platform for analysing GNSS

systems over a service volume at specific moments

in time, with the capability to then inspect at specific

locations within that service volume.

For the performance assessment, the strategy is

that the satellite geometry and UDRE values are

used to compute the accuracy and integrity

(protection level) values at a grid of analysis points

within the defined coverage area for a number of

different times.

There are defined minimum maritime user

requirements for many different applications from

IMO (2001). Generally, the operations can be

grouped into high (0.1m accuracy, 0.25m integrity),

medium (1m, 2.5m) and low (10m, 25m)

accuracy/integrity requirements.

We can compare the performance results against

the required navigation performance to see if the

option has the potential to provide an adequate

service.

3.2 Summary of L1 DGPS Performance

From the results of the L1 DGPS performance

assessment the following conclusions can be drawn:

− All the code positioning options under considera-

tion for 1st generation re-capitalization are very

close to meeting the low accuracy/integrity

requirements for horizontal performance with L1

GPS data. The simulation results show that the

horizontal accuracy can meet the 10m require-

ment 100% of the time for coastal areas and the

horizontal integrity meets the 25m limit about 95-

99% of the time;

− For the new 2nd generation re-capitalization

options, then the SW Radio + SW RSIM option is

the only one that can match the 1st generation

options in terms of horizontal performance for

legacy users (L1 GPS);

− The vertical performance is slightly worse than

the horizontal performance for all options. This is

a function of the satellite geometry and is

unavoidable when using only GNSS

measurements. However, the vertical

performance requirements only apply to certain

specialised applications and so the majority of

users’ needs are met;

− None of the code positioning options are close to

meeting the medium accuracy/integrity

requirements;

− The OSNET VRS (carrier phase) option is the

only one that is even close to meeting the high

accuracy/integrity requirements, although the area

where the service is available is limited to the

area within the reference station network and up

to 20 km outside of it (Cruddace 2005).

3.3 Summary of L1/L5 DGNSS Performance

For the results of the L1/L5 DGNSS performance

assessment the following conclusions can be drawn:

− All the code positioning options from both 1st

generation and new 2nd generation options can

easily meet the low accuracy/integrity

requirements;

− All the 1st generation code positioning options

are very close to meeting the medium accuracy

/integrity requirements in the horizontal domain.

The simulations show that the 1m accuracy

requirement can be met 99.9% of the time in the

majority of coastal locations, except for a few

areas around the Irish coast. The horizontal

integrity can meet the 2.5 m requirements around

95% of the time at all locations;

− There is the possibility that the OSNET VRS

code option could improve the performance in

certain areas by changing the locations of the

virtual reference stations, without having to move

the transmitters;

− For the new 2nd generation re-capitalization

options, then the SW Radio + SW RSIM option is

the only one that can match the 1st generation

options in terms of horizontal performance;

− The vertical positioning/integrity performance is

worse than the horizontal performance for all

options due to the satellite geometry. However,

the vertical performance requirements only apply

to certain specialised applications and so the

majority of users’ needs are met;

− The OSNET VRS (carrier phase) option performs

better than the EGNOS WARTK solution and is

very close to meeting the high accuracy/integrity

horizontal requirements. However, the area where

the service is available is limited to the area

within the reference station network and up to 20

km outside (Cruddace 2005) whereas the EGNOS

WARTK service would be available at all

locations within the area served by the GLA.

4 RISK ANALYSIS

4.1 1

Generation Options

For 1st generation re-capitalization then there are 3

DGPS code options and one high accuracy option.

The high accuracy option (OSNET VRS RTK) is an

extension of the OSNET VRS code option but they

are not interlinked, and so it is possible to operate

the OSNET VRS code service without the OSNET

VRS RTK service, and vice versa. Therefore the

risks are considered separately, although many of

them are common.

For the 3 DGPS code options (HW RSIM, SW

RSIM and OSNET VRS code), the following results

are apparent:

− Considering only those risks relevant to the 1st

generation re-capitalization, the option with the

highest technical risk is the OSNET VRS code

option. This is due to the fact that the architecture

is quite different from the existing DGPS service

and the SW is more complex and so there is a

slightly greater probability for this option that it

will not be available in time for 1st generation;

− The option with the highest operational risk is

again the OSNET VRS code option. This is due

to the fact that this option relies on 3rd party data

input, and any timeliness or reliability issues with

the data will affect the ability of the solution to

provide correction messages with correct

accuracy and integrity. As there is a single

processing facility that computes corrections for

all sites, any complete interruption to the OS data

means that messages from the whole DGPS

transmitter network are affected;

− The HW RSIM and SW RSIM options have the

same technical risk. However, the SW RSIM

option has a higher operational risk related to the

fact that an upgrade of the operating system (e.g.

from Windows NT to XP) may cause

incompatibilities and affect the availability of the

service. Although the HW RSIM option has the

same risk, the consequence is lower because this

option is based on HW and will not be affected so

much;

− Overall, the option with the least risk for 1st

generation re-capitalization is the HW RSIM

option. The OSNET VRS code option is by far

the riskiest option with several risks that have a

high total score;

− When considering the risks associated with

upgrade of these options for the 2nd generation

programme, the option with the highest technical

risk is the OSNET VRS code option. This is due

to the fact that this option is reliant on 3rd party

enhancement of the reference receiver network so

that L1/L5 GNSS measurements are available;

− For upgrade to 2nd generation then the option

with the lowest risk is the SW RSIM option. This

is because there is potentially greater cost

involved in upgrading the HW of the HW RSIM

option to meet 2nd generation requirements

compared to the upgrade for the SW RSIM

option.

The OSNET VRS RTK option is the only one of

those considered that could potentially be used for a

high accuracy service. However, it has significant

risk, especially on the technical side:

− The high technical risk is due to the fact that the

high accuracy solution is limited to the area

within, and up to 20km outside of, the reference

network (Cruddace 2005). This means that the

option cannot provide a high accuracy service at

all locations within the coastal region up to 50

nautical miles from the coast;

− For the operational risk then the highest score is

due to the fact that the existing transmitters

cannot be used because their bandwidth is not

great enough and so a different broadcast method

has to be used. The actual method is not

consolidated but some of the options, e.g. GSM,

may not provide the required level of message

availability;

− When considering the risks associated with

upgrade for the 2nd generation re-capitalization

programme, the OSNET VRS RTK option has a

high risk due to the fact that this option is reliant

on 3rd party enhancement of the reference

receivers so that L1/L5 GNSS measurements are

available.

Such high risks for the OSNET VRS RTK option

make it unsuitable for implementation and so this

option is not considered further in this study.

4.2 New 2

Generation Options

Although the new 2nd generation re-capitalization

options are not considered for 1st generation re-

capitalization, they should be able to provide an L1

DGPS service to cater for legacy users. There are 3

options for providing a DGPS code service and 1

option that could potentially be used for a high

accuracy service (EGNOS WARTK).

For the new 2nd generation DGPS code options,

the remaining risks after mitigation should be

considered. The following results are apparent:

− The option with the highest technical risk is the

EGNOS RTCA option. This is due to the fact that

the messages are not compatible with RTCM

version 2.X and so legacy users cannot be

supported with this option alone;

− The EGNOS RTCA and EGNOS RTCM options

also have high risk due to their degraded

performance compared to the other DGPS code

options;

− The EGNOS RTCA and RTCM options also have

a high technical risk due to their reliance on a 3rd

party enhancement programme. At the present

time, EGNOS is almost operational to provide

corrections for GPS L1. In order to be considered

for 2nd generation, EGNOS must have been

enhanced so that it covers both GPS and Galileo

on L1/L5, and also the EDAS must be

operational. There are evolution plans for

EGNOS, but until the plans are implemented it

remains a risk that the products will not be

available in time;

− The SW Radio + SW RSIM option has the lowest

technical risk of all the new 2nd generation DGPS

code options. However, the risk for this option is

still slightly higher than for the upgrading of HW

and SW RSIM options from 1st generation;

− All options have an equally high operational risk

due to operating system upgrade, and the fact that

any upgrade result in incompatibilities and impact

availability. However, the EGNOS RTCA and

EGNOS RTCM options have an additional

operational risk due to reliance on 3rd party input

data.

For these points, the main finding is that the

EGNOS RTCA should not be considered further in

this study because it does not support legacy users,

and so will not allow the GLA to fulfil all their core

obligations.

For the high accuracy service option (EGNOS

WARTK) then the following points are observed:

− EGNOS WARTK has a very high technical risk

because the solution relies on dual frequency data

at both the system and user level, and so it is not

possible to provide a service for legacy users;

− The EGNOS WARTK option also has a high

technical risk due to the unavailability of suitable

processing SW for purchase at the current time. In

fact, this approach has only been demonstrated in

research papers and so there is little evidence that

a feasible service can be provided;

− It is also the case that the existing DGPS

transmitter network cannot be used for

broadcasting the EGNOS WARTK messages

because the bandwidth is not sufficient. Therefore

alternative transmission means will have to be

found.

Because the EGNOS WARTK option has a high

technical risk and cannot support legacy users, this

option is not considered further. Unfortunately this

means that both high accuracy options (OSNET

VRS RTK and EGNOS WARTK) have been

discounted due to excessive risk.

5 COSTS

The aim of the economic analysis is to provide the

estimated costs for the different re-capitalization

options. Following the risk analysis, the remaining

candidate options are as follows:

− 1st generation re-capitalization options

− Procurement of Upgraded DGPS equipment

− Procurement of upgraded HW in combination

with SW RSIM

− OSNET VRS (code)

− New 2nd generation re-capitalization options

− SW Radio in combination with SW RSIM

− EGNOS RTCM

It is not the aim of this task to determine the exact

final costs for every single piece of HW and SW and

all running costs associated with the options. Rather

it is expected that this analysis will give ball-park

figures for each of the options in order to allow a

simple comparison of order of magnitude costs.

Only the initial costs have been included at this

stage. Whole-life costs, including running costs and

decommissioning would give a better assessment,

preferably taking account of the interaction between

and 2

generation options. However, due to

uncertainties and lack of data on many of the

options, this was considered out of the scope of this

study.

From analysing the final costs for each option, the

following points are observed:

− For 1st generation re-capitalization,

− The option with lowest procurement costs is

the VRS with OSNET option. This is because

the option operates on a central Processing

Facility architecture and so the costs are

reduced compared to the separate RSIM site

options;

− However, the operational costs associated with

the VRS with OSNET option are currently

unknown because the information on data

access costs to the OS reference data has not

yet been provided. This increases risk for this

option as the full costs cannot be estimated;

− Disregarding the VRS option, the baseline HW

RSIM option has lowest cost;

− The SW RSIM option is significantly more

expensive than the Procurement of upgraded

DGPS equipment (HW) option. This is because

any saving in GNSS Rx cost (because of

reduced functionality) is more than wiped out

by the additional RSIM SW costs;

− For both the HW and SW RSIM options,

planning for easy upgrade to 2nd generation

requirements at this stage adds significant extra

cost to the procured GNSS equipment

− For 2nd generation re-capitalization;

− The cheapest option is the HW RSIM option;

− The EGNOS RTCM option offers some saving

over the SW RSIM option because the

reference station SW is less complex. Further

savings could be achieved if the EGNOS

RTCM option moved to a central Processing

Facility approach (like the VRS option) but this

puts extra burden on the comms and would

need further study;

− The operational costs for the EGNOS RTCM

option are not confirmed as information on

costs to access the EGNOS messages via

EDAS is not available. It is assumed to be free

but this may not be the case and so this

therefore increases risk for this option as the

full costs cannot be estimated;

− The most expensive option is the SW Radio +

SW RSIM option. However, due to the

immaturity of the technology, and the long time

into the future we are looking at, the costs

estimates may be unreliable.

6 CONCLUSIONS AND RECOMMENDATIONS

The lowest cost, lowest risk option would be to

replace existing hardware with similar, dedicated

RSIM based on known, commercially available

technology. However, the choice of suppliers is

limited and once chosen, it may be difficult to

change it to meet emerging requirements.

The flexibility provided by the SW RSIM option

could overcome this problem and it should not be

ruled out at this stage. Therefore the following

recommendations are proposed for the DGPS re-

capitalization:

1 The Hardware RSIM option should be adopted

for the 1st Generation Re-capitalization, on

grounds of lowest cost and risk;

2 Transition to the SW RSIM option during the

lifetime of the 1st Generation system should be

considered, if suitable proven software becomes

available at reasonable cost;

3 Study of the 2nd Generation options should be a

continuing project, running in parallel with the 1st

Generation Re-capitalization work.

REFERENCES

Cruddace, P 2005. Enabling Application Development through

a Supporting National GPS Infrastructure. Nav 05, ‘Pushing

the Boundaries’, London, 1-3 November.

Hernandez-Pajeras, Juan, Sanz, Orus, Garcia-Rodgriguez,

Colombo 2004. Wide Area Real Time Kinematics with

Galileo and GPS Signals. ION GNSS 17th International

Technical Meeting of the Satellite Division, Long Beach,

CA, 21-24 September.

IMO, 2001. Resolution A.915(22) Revised Maritime Policy

and Requirements for a Future Global Navigation Satellite

System (GNSS). 22nd Session of the IMO Assembly. 19-30

November.

Wolfe, D 2006. Next Generation Differential GPS

Architecture. IALA DGNSS, London, 13-16 March.