The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008 
* GPS Antenna Stability 
■ Multi-path 
■ Easy access 
■ Network Geometry 
■ Electrical Power and Communications for Real Time 
Kinematic (RTK) Surveys 
■ Equipment Protection 
According to the island topography, the distance between 
stations of the network must be under 35 Km, while the height 
could vary from 100 m up to 1900 m. 
In addition to the aforementioned key factors, other important 
details like: 
■ Monument construction 
■ Antenna installation 
■ Water proofing 
should be taken into account. 
Special attention should be given to the radio noise sources and 
in general the interference which can be assumed as originated 
from these sources. As it is known, the strength of the GPS 
satellite signal is very low, so nearby sources of electrical or 
radio noise can cause significant problems by interfering with 
the GPS signals. These types of noise may originate from: 
■ Electrical transmission lines 
■ Nearby commercial radio or television broadcast stations 
■ Radio dispatch stations 
■ Police, fire & other emergency services 
■ Taxi services 
■ Pick up and delivery services. 
■ Airports 
All these parameters and restrictions have been taken into 
account while designing the PSI network and quality of signal 
tests were performed (Leica 2006). The designed network 
geometry is depicted in Fig.l 
3. NETWORK ASSESSMENT 
3.1 The Challenge of VRS 
Recent developments in differential GPS services are focusing 
mainly on the reduction of the number of permanent reference 
stations, required to cover a certain area, and on the extension 
of baseline lengths between reference and rover receivers. The 
most advanced technique nowadays is the Virtual Reference 
Station networks. The name of this method results from the fact 
that artificial observations for a (non existing) ‘virtual’ GPS 
station are created, using the real observations of a multiple 
reference station network. 
The concept idea is that a user is activated inside a network of 
permanent reference stations, that are connected with a central 
control station. User receiver transmits its navigated position to 
the central station. The communication is usually performed 
with cellular phones (GSM) and in future with Universal 
Mobile Telecommunication Service (Hada et.al, 1999,2000). 
The contribution of the network is the knowledge of the errors 
behaviour, specifically of those that are distance dependent. So, 
ionospheric and tropospheric refraction and orbit errors are 
modelled and can be interpolated at any position inside the 
network. The artificial observations for the user approximate 
location are created and then transferred (through RTCM 
format) back to the user (rover station). On the rover side, 
standard RTK or DGPS algorithms are used to obtain the 
correct position. The result is the increment of baseline lengths 
and the reduction of initialization times (Rizos, 2002, Fotiou 
and Pikridas 2006). 
During these projects, the PSI network was established and the 
VRS method was tested with C/A code data using custom- 
developed software. The results showed always a positional 
accuracy of better than 1 m. 
3.2 Description of Applications 
In the next paragraphs, a number of test cases, based on the PSI 
network, are presented. These span a wide spectrum of different 
geomatics problems, while the main focus remains the proof of 
the productivity of the network and the evaluation of its 
operation. 
For this purpose, different techniques, hardware and software 
components have been used. The applications cover also a wide 
range of accuracy requirements (from several cm to several m) 
and all of them have been performed in Cyprus. 
3.2.1 Case 1: Application of VRS Technique 
In order to test the accuracy achievements of the virtual 
reference station concept, a test network, consisting of three 
dual frequency GPS receivers, was temporary established in the 
broader area of Pafos. The network was first measured and 
adjusted in order to determine accurate coordinates. Inside the 
area (triangle) of the three permanent stations a low cost (less 
than 3.000 €) rover receiver was also activated to perform a 
kinematic chain of about 2 hour’s duration. The survey took 
place almost 20 Km away from the closest station. A user 
friendly VRS software called “3VRS” was developed and 
artificial C/A code observations were created for the first 
navigated position of the rover receiver using the data of the 
closest station. Investigating the effectiveness of the VRS 
technique three different solutions were derived. 
■ Phase solution with fix ambiguities 
■ DGPS solution with post-processing 
■ DGPS solution using VRS data 
The results from the comparisons between the various solutions 
show a precision for the X and Y components better then 0.5 m 
and about 1.5 times worse for the Z component (Table 1). 
Therefore our tests prove that an accuracy increment for this 
kind of baseline lengths can be obtained when the VRS method 
is applied. 
Difference of 
components 
Maximum 
value (m) 
Minimum 
value (m) 
Mean 
value (m) 
AX 
0.62 
0.14 
0.26 
AY 
0.74 
0.16 
0.32 
AZ 
0.84 
0.01 
0.65 
AS (position) 
1.28 
0.21 
0.75 
Table 1. Characteristic values of the differences of the 
component and position between the VRS (DGPS) solution 
(C/A code and LI frequency processed only) and the one 
derived from the phase measurements 
326