The software sends to the rover a correction term that is the sum 
of the atmospheric delays and of the radial component of the 
ephemeris error (difference between predict and broadcast 
ephemerides). 
The figure (1) represent the RTCM PRC and the one way 
estimated atmospheric correction for the satellite 24. There are 
differences up to 3 meters (and more for some other satellite), 
mainly due to the ephemeris error. 
  
  
  
  
  
  
PRN24 PRC 
m fr 
ARA RTCM PRC 
= + ^ + One Ner r. 
2 e | IONO«TROPO | 
E 
4 in 
0 500 1200 1800 2400 3000 3600 
t{soh] 
Figure 1. RTCM PRC and one way estimated atmospheric 
correction (Iono + Tropo) for the PRN24. Note that 
in t — 1800-2100 soh the PRC is misestimated due 
to constellation changing. 
TORI PRC 
  
PRC [m] 
  
  
  
  
t[sow] 
Figure 2. One way ALARIS estimated atmospheric correction 
(lono + Tropo) for all the satellite in view. 
4.3 Test on RTK algorithm 
We have performed a test on RTK positioning over short and 
long baselines using the software ALARIS. The rover receiver 
moves on a cart on a rectilinear trajectory, after a static 
initialization of 15 minutes. Then stops for other 15 minutes. 
The trajectory of the cart is tracked in post processing using a 
reference station 150 m far. This post processed trajectory is the 
“truth” of the test, and is compared with the trajectories 
generated in RTK mode over long baselines. 
The reference stations used in the test are: 
eo Torino TORI baseline — 150m 
e Biella BIET. baseline ~ 60 km 
e Novara NOVA baseline ~ 80 km 
e Vigevano VIGE baseline — 100 km 
   
80 
  
Figure 3. The rover receiver and antenna mounted on the cart. 
  
Em) 
Figure 4. The trajectory of the cart tracked in simulated RTK 
using the reference station of Vigevano (VIGE) is 
represented by the dotted line. The baseline is about 
100 km. The PPK solution on the short baseline, that 
is the “truth” of the test, is represented by the 
continuous line. 
  
  
  
  
  
  
  
  
  
- 
& dp, 
> " om veneni 
e 4 
3 ; 
£ Aat di direi de ram itat irt "i 
i Vagina va , à 
3 
= 
= 
ü 
  
^ i eu 
  
  
tis) 
Planimetric offsets in the single frequency solution 
performed on the baseline of 150 m. The upper 
series is the error in East, the lower the error in 
North. 
Figure 5. 
uv € 
  
Figure 6. 
Over lon; 
bad amb 
residual t 
compute | 
= 
at bel 
  
Figure 7. 
44 Test 
The goal 
cost rece 
engineeri 
We have 
4000SSi) 
eMap) co 
observati 
Garmin r 
of the se 
provided