to multipath compared to a 6 channel receiver. 
Also, differences in the receiver tracking loop 
design may also affect the performance of a 
particular receiver in a sub-optimal environment 
since this may affect both multipath 
susceptibility and satellite re-acquisition times. 
Several tests have been conducted to assess the 
performance of GPS in urban environments, and one 
such test is reported in Melgard et al. (1994). In 
this case, three receivers were assessed, i.e. two six 
channel receivers and one ten channel receiver, in 
both a residential and urban environment. The 
three receivers were mounted in a truck which 
drove a trajectory consisting of both residential 
and urban areas. Three test runs were conducted, 
each lasting about 1.5 hours. The results were 
analysed according the availability of GPS, 
which is defined as the percentage of time that at 
least four satellites were observed. Table 4 gives 
the statistics for each of the three receiver types. 
The table clearly shows that the performance from 
each receiver is not equal, and that the 10 channel 
receiver has better availability statistics than 
the six channel receivers, as expected. The 
statistics for the entire trajectory are about 9096 for 
the 10 channel receiver while it is about 50-7096 
for the six channel receivers. 
Table 4 also shows that the performance is 
significantly better in a residential area as 
compared to the downtown area. The availability 
drops to 50-70% from 92-100 % for the ten channel 
receiver and 16-43% for the six channel receivers 
from 23-91%. These tests confirm that receiver 
selection is critical for cases when sub-optimal 
conditions will be observed. 
Table 4: GPS Availability Statistics (%) 
  
  
  
  
Test# Trajectory Receiver 
1 2 3 
Downtown 51 16 40 
1 Residential 98 91 78 
Entire Traj. 88 67 73 
Downtown 69 39 43 
2 Residential 100 68 61 
Entire Traj. 92 66 75 
Downtown 57 31 35 
3 Residential 92 24 23 
Entire Traj. 88 51 60 
  
4.3 Precision Farming 
The application of GPS to the agricultural 
industry is beginning to accelerate due to the 
168 
reduction in receiver costs. The University of 
Calgary, together with Alberta Agriculture, is 
currently developing a system which can be used to 
measure the variability of crop production as a 
function of location (Gehue et al., 1994). The 
system consists of several components which 
include GPS receivers for positioning, a yield 
monitoring system which outputs the instantaneous 
Bu per acre, an EM conductivity meter for salinity 
measurements, and soil samples for determination 
of soil types and nutrients. The DGPS/yield 
monitoring data can be collected under normal 
combining operations. Two 10-channel C/A code 
narrow correlator spacing NovAtel GPSCard™ 
sensors were used as they have shown to provide 
sub-metre accuracy in previous field tests using a 
robust carrier phase smoothing of the code 
approach (e.g. Cannon and Lachapelle, 1992). Once 
all this data has been collected and analysed, 
variable fertilizing can be used in subsequent years 
to optimize productivity. DGPS can be used as a 
real-time navigation and guidance tool for this 
application. 
A GIS is used to organize and collate the variety of 
data collected from year to year. Various layers of 
information can be draped (using surfacing 
routines) over each other through the DGPS 
position. It is then possible to overlay crop yields 
with salinity, for example, and draw conclusions 
based on the relationship (i.e., high salinity 
results in low yield). Further, predictions can be 
made on how to handle these regions when it is 
time to fertilize and/or seed. The GIS can also be 
updated with external information such as aerial 
photography or remote sensing. Current 
information layers of topography, salinity, soil 
type, nutrients, and crop yields were collected for 
the 1993 harvest. 
The goal of the precision farming project is to 
optimize the yield-input relationship for any 
given field. Since this is the first harvest season of 
the project, results to date only give an indication 
of the variation within the field for this year, 
with the condition that it has been treated 
equally in the past. A surfacing routine is used to 
view the yield variations with the result shown in 
Figure 5 for a subset of the test site. The figure 
clearly shows that the field is not homogeneous in 
yield. Surface analysis of crop yields allows 
agricultural researchers to identify problem areas 
and to maintain data quality by detecting results 
such as the sudden peak in the NE corner of the 
field which is most likely caused by high 
moisture in the swath which returns a false 
reading from the yield sensor. 
  
Fig. 5 
Once all : 
their res| 
between 
informati 
potential 
specific s 
have sev 
have di 
fertilizers 
yield. Th 
take into 
topograp 
applicati. 
The test f 
steep noi 
monitor : 
the movii 
km from 
harvestec 
Novembe 
locations 
Table 5: 
Smoothir 
  
Dat 
  
| Sept. 2 
Nov. 9 
  
  
The posi: 
project are 
Two techr 
namely a 
and a on- 
(OTF). T 
verified 
using the: 
Was used